Methylglyoxal Down-Regulates the Expression of Cell Cycle Associated
Total Page：16
File Type：pdf, Size：1020Kb
www.nature.com/scientificreports OPEN Methylglyoxal down-regulates the expression of cell cycle associated genes and activates the p53 Received: 8 May 2018 Accepted: 12 December 2018 pathway in human umbilical vein Published: xx xx xxxx endothelial cells Jana D. Braun1, Diego O. Pastene1, Annette Breedijk1, Angelica Rodriguez1, Björn B. Hofmann1, Carsten Sticht2, Elke von Ochsenstein3, Heike Allgayer3, Jacob van den Born4, Stephan Bakker4, Sibylle J. Hauske1, Bernhard K. Krämer1, Benito A. Yard1 & Thomas Albrecht1 Although methylglyoxal (MGO) has emerged as key mediator of diabetic microvascular complications, the infuence of MGO on the vascular transcriptome has not thoroughly been assessed. Since diabetes is associated with low grade infammation causing sustained nuclear factor-kappa B (NF-κB) activation, the current study addressed 1) to what extent MGO changes the transcriptome of human umbilical vein endothelial cells (HUVECs) exposed to an infammatory milieu, 2) what are the dominant pathways by which these changes occur and 3) to what extent is this afected by carnosine, a putative scavenger of MGO. Microarray analysis revealed that exposure of HUVECs to high MGO concentrations signifcantly changes gene expression, characterized by prominent down-regulation of cell cycle associated genes and up-regulation of heme oxygenase-1 (HO-1). KEGG-based pathway analysis identifed six signifcantly enriched pathways of which the p53 pathway was the most afected. No signifcant enrichment of infammatory pathways was found, yet, MGO did inhibit VCAM-1 expression in Western blot analysis. Carnosine signifcantly counteracted MGO-mediated changes in a subset of diferentially expressed genes. Collectively, our results suggest that MGO initiates distinct transcriptional changes in cell cycle/apoptosis genes, which may explain MGO toxicity at high concentrations. MGO did not augment TNF-α induced infammation. Te incidence of diabetes is rapidly increasing to epidemic proportions, afecting by 2040 1 out of 10 persons globally according to recent estimates1. Because diabetes is associated with hyperglycemia-specifc micro- and macro-vascular complications, e.g. diabetic nephropathy (DN) and cardiovascular disease, the rapid increase of numbers of people with diabetes will augment the economic costs for morbidity and mortality in coming years thereby absorbing a considerable proportion of the healthcare budget. For decades, hyperglycemia was considered to be the main driver of late diabetic complications and as such the primary therapeutic target in diabetic patients. Large trials assessing the efect of intensive glycemic control in the general diabetic population2,3 have indeed suggested that tighter glycemic control may improve microvas- cular outcomes in patients with diabetes, yet, the relationship between intensive glycemic control and reduced incidence and/or progression of macro-vascular complications is less clear4,5. Even though our understanding of micro- and macro-vascular complications has signifcantly improved, the therapeutic options for diabetic patients 1Department of Nephrology, Endocrinology and Rheumatology, Fifth Department of Medicine, Medical Faculty Mannheim, University of Heidelberg, Mannheim, Germany. 2Center of Medical Research, Medical Faculty Mannheim, University of Heidelberg, Mannheim, Germany. 3Department of Experimental Surgery - Cancer Metastasis, Medical Faculty Mannheim, University of Heidelberg, Mannheim, Germany. 4Department of Internal Medicine, University Medical Centre Groningen, Groningen, Netherlands. Correspondence and requests for materials should be addressed to J.D.B. (email: [email protected]) SCIENTIFIC REPORTS | (2019) 9:1152 | https://doi.org/10.1038/s41598-018-37937-1 1 www.nature.com/scientificreports/ are mostly still limited to blood pressure control, hyperglycemia management, use of a statin and reduction of proteinuria via renin-angiotensin blockade. New therapeutic developments such as SGLT-2 inhibition and GLP-1 agonistic agents, that have recently been shown to improve proteinuria, hold promise to reduce the medical and economic burden associated with DN6–8. Te role of oxidative stress as a causal link in the development of hyperglycemia-associated complications has been highlighted in many studies9,10. Oxidative stress may cause protein modifcations, either directly via reactive oxygen species (ROS), or indirectly by reactive carbonyl products formed by auto-oxidation of carbo- hydrates, lipids or amino acids. While auto-oxidation of carbohydrates yields precursors of advanced glycation end-products (AGE), e.g. glyoxal, methylglyoxal (MGO) and glycolaldehydes, lipid peroxidation also generates precursors of advanced lipoxidation end-product (ALE), e.g. malondialdehyde (MDA) and 4-hydroxynonenal (4-HNE)11,12. AGE and ALE can evoke a variety of biological responses, e.g. stimulation of extracellular matrix production, induction of infammatory responses and inhibition of proliferation, all of which may perpetuate the progression of diabetic lesions to various degrees13,14. Amongst the precursors of AGE, MGO is a potent glycating agent by far more reactive compared to glu- cose15. It has been suggested that MGO covalently modifes the 20S proteasome16 thereby decreasing the abil- ity of diabetic kidneys to eliminate malfunctioning or damaged proteins17. Compatible with this suggestion is the fnding that knockdown of glyoxalase-1 in non-diabetic mice results in renal lesions indistinguishable from those of diabetic mice, while overexpression of glyoxalase-1 in diabetic mice prevents the development of nephropathy18. Other studies have shown that MGO impairs HIF-1α degradation and signaling19,20 and activates AMPK mediated autophagic degradation of thioredoxin 121, thus emphasizing its infuence on redox homeo- stasis22. Despite the clear association between reactive carbonyl species and diabetic complications, their mode of action on endothelial cells is discussed ambiguously23–27. A general fnding throughout all studies is however that MGO causes endothelial damage, albeit that diferent MGO concentrations have been reported at which this occurs23,28–30. It is believed that endothelial damage results from apoptosis, yet a comprehensive pathway analysis to our knowledge has not been reported. MGO-mediated apoptosis can be prevented by glycation end-product inhibitors31,32, by anti-oxidants33,34 and interestingly by cPLA2 inhibition35. In the latter study, it also has been suggested that MGO inhibits phosphorylation of nuclear factor-κB (NF-κB) and that pharmacological inhibition of NF-κB further increases MGO-induced apoptosis of human umbilical vein endothelial cells (HUVECs). For a better understanding of MGO-induced cytotoxicity, we assessed to what extent MGO changes the transcriptome of HUVECs exposed to a concurrent infammatory milieu. In addition, we assessed to what extent this is afected by carnosine (CN), a histidine containing dipeptide with reactive carbonyl scavenging properties. Results Methylglyoxal signifcantly alters the gene expression profle. To assess the infuence of MGO on gene expression in HUVECs that are exposed to the pro-infammatory cytokine TNF-α, large-scale gene expres- sion profling was performed. We frst determined susceptibility of HUVECs to MGO by assessing cell viability over a wide range of MGO concentrations (0–3.2 mM). As depicted in Fig. 1, MGO did not change cell morphol- ogy at 800 µM while at 1.6 mM cells started to detach accompanied by 7-AAD positive staining in FACS. Since at 800 µM of MGO no obvious toxicity was noted, we used this and a twofold lower MGO concentra- tion for subsequent gene expression profling. Based on this, in microarray analysis six experimental groups of TNF-α exposed HUVECs were established as follows: no MGO ± CN, 400 µM MGO ± CN and 800 µM MGO ± CN (24 hours incubation). Whereas 400 µM of MGO did not signifcantly change gene expression com- pared to no MGO, the transcriptional profle was strongly afected by 800 µM of MGO. Te p-value distribu- tion on the efect of 800 µM of MGO revealed 4.6 times more genes to be diferentially expressed as would be expected if the null hypothesis was true (p < 0.0001 by one-tailed binomial test) (Fig. 2A). By applying an adjusted p-value < 0.05 (adjusted for multiple testing) and a fold change (FC) threshold of ≥1.5, a total of 855 transcripts (342 up-regulated and 513 down-regulated) were found to be diferentially expressed as depicted by the corre- sponding volcano plot (Fig. 2B). Te ten most down-regulated, respectively up-regulated genes are enlisted in Table 1. As already suggested by the skewing of the volcano plot to the lef (more genes, higher signifcance and larger FC), the MGO mediated change in gene expression profle was dominated more by down-regulation than by up-regulation of gene expres- sion. Amongst the up-regulated genes heme oxygenase-1 (HO-1) was ranked the highest with an approximately 3.8 fold increased expression (p = 9.63E-5). Down-regulated genes mostly involved cell cycle dependent genes, e.g. topoisomerase (DNA) II alpha (TOP2A), abnormal spindle microtubule assembly (ASPM), kinesin family member 20 A (KIF20A), marker of proliferation Ki-67 (MKI67) and cyclin A2 (CCNA2) (all p < 0.001) (Table 1). Enriched genes cluster in the categories cell cycle, mitosis and apoptosis. All 855 diferentially expressed genes (DEGs) were subjected to Gene Ontology (GO) analysis. In the GO Biological Process category, the top fve overrepresented annotations were cell cycle